We have developed a multichannel taste sensor with artifical lipid membranes and have applied it to quantification of taste of green tea. We used multiple regression analysis and found high correlations of outputs of the taste sensor with the results of sensory test (taste, flavor and color) and chemical analyses (amino acids and tannin that are main taste substances in green tea). It is concluded that the taste sensor has a potential for quantification of taste of green tea. The taste sensor responds not only to amino acids and tannin, but also to many other taste substances, and hence it contains much more taste information than conventional chemical analyses.